RuntimeError: (PreconditionNotMet) The third-party dynamic library (cudnn64_8.dll) that Paddle depends on is not configured correctly. (error code is 126)

[gtbloody]

运行paddleocr --image_dir E:\PaddleOCR\img1.pdf --use_angle_cls true --use_gpu true的时候报错，这是找不到cudnn64_8.dll？

RuntimeError: (PreconditionNotMet) The third-party dynamic library (cudnn64_8.dll) that Paddle depends on is not configured correctly. (error code is 126)
Suggestions:

1. Check if the third-party dynamic library (e.g. CUDA, CUDNN) is installed correctly and its version is matched with paddlepaddle you installed.
2. Configure third-party dynamic library environment variables as follows:

• Linux: set LD_LIBRARY_PATH by export LD_LIBRARY_PATH=...
• Windows: set PATH by `set PATH=XXX; (at ..\paddle\phi\backends\dynload\dynamic_loader.cc:301)

————————————————————————————
以上是报错，以下是我电脑和环境的情况，cuda版本11.8，用python -m pip install paddlepaddle-gpu==2.5.2 -i https://pypi.tuna.tsinghua.edu.cn/simple 安装的paddlepaddle-gpu，用pip install paddleocr安装的paddleocr2.7.0.3
————————————————————————————

(base) (venv) PS E:\PaddleOCR> pip list
Package Version

---

albumentations 1.3.1
anyio 4.0.0
astor 0.8.1
attrdict 2.0.1
Babel 2.13.1
bce-python-sdk 0.8.95
beautifulsoup4 4.12.2
blinker 1.7.0
cachetools 5.3.2
certifi 2023.7.22
charset-normalizer 3.2.0
click 8.1.7
colorama 0.4.6
contourpy 1.1.0
cssselect 1.2.0
cssutils 2.9.0
cycler 0.11.0
Cython 3.0.5
decorator 5.1.1
et-xmlfile 1.1.0
exceptiongroup 1.1.3
filelock 3.12.2
fire 0.5.0
Flask 3.0.0
flask-babel 4.0.0
fonttools 4.42.0
future 0.18.3
glog 0.3.1
h11 0.14.0
httpcore 1.0.1
httpx 0.25.1
idna 3.4
imageio 2.31.1
imgaug 0.4.0
itsdangerous 2.1.2
Jinja2 3.1.2
joblib 1.3.2
kiwisolver 1.4.4
lmdb 1.4.1
lxml 4.9.3
MarkupSafe 2.1.3
matplotlib 3.7.2
mpmath 1.3.0
munch 4.0.0
networkx 3.1
numpy 1.25.2
opencv-contrib-python 4.6.0.66
opencv-python 4.6.0.66
opencv-python-headless 4.8.0.76
openpyxl 3.1.2
opt-einsum 3.3.0
packaging 23.1
paddleocr 2.7.0.3
paddlepaddle-gpu 2.5.2
pandas 2.1.2
pdf2docx 0.5.6
Pillow 10.0.0
pip 23.3.1
premailer 3.10.0
pretrainedmodels 0.7.4
protobuf 3.20.2
psutil 5.9.6
pyclipper 1.3.0.post5
pycryptodome 3.19.0
PyMuPDF 1.20.2
pyparsing 3.0.9
python-dateutil 2.8.2
python-docx 1.1.0
python-gflags 3.1.2
pytz 2023.3.post1
PyWavelets 1.4.1
PyYAML 6.0.1
qudida 0.0.4
rapidfuzz 3.5.2
rarfile 4.1
requests 2.31.0
scikit-image 0.18.1
scikit-learn 1.3.0
scipy 1.11.1
setuptools 68.2.0
shapely 2.0.2
six 1.16.0
sniffio 1.3.0
soupsieve 2.5
sympy 1.12
tensorboardX 2.6.2
termcolor 2.3.0
threadpoolctl 3.2.0
tifffile 2023.8.12
torch 2.0.1+cu118
torchaudio 2.0.2+cu118
torchsummary 1.5.1
torchvision 0.15.2+cu118
tqdm 4.66.1
typing_extensions 4.7.1
tzdata 2023.3
urllib3 2.0.4
visualdl 2.5.3
Werkzeug 3.0.1
wheel 0.41.2
(base) (venv) PS E:\PaddleOCR> nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:41:10_Pacific_Daylight_Time_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0

[gtbloody]

解决了，是我没装cudnn，把cudnn压缩包里bin目录下的cudnn64_8.dll等7个文件解压到C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin里就好了

[psgetit]

在实际应用过程中可能受多种影响缺的不只单个动态链接库，可以看下这个https://blog.csdn.net/shuaishishi/article/details/136619380?spm=1001.2014.3001.5501

[AlphaBoom]

解决了，是我没装cudnn，把cudnn压缩包里bin目录下的cudnn64_8.dll等7个文件解压到C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.8\bin里就好了

现在下载下来的v9版本后缀都是_9，把cudnn64_9.dll复制一份名称改成cudnn64_8.dll也能用。

[wisecsj]

paddlepaddle-gpu目前(2024.6)应该只支持到cudnn v8的大版本。但是官方下载下来的都是v9了，所以会报该错。去cudnn archive下v8的就好。

[Myles091]

为啥我下载下来的后缀时_8，也粘贴进去了，为啥还是不行，我干

[hopesala]

复制完，9改成8，我的报错消失了，可以运行了

[freemedom]

如果安了torch的话也可以用torch的库，...\site-packages\torch\lib\